Yordanova 2026 — Corneal CFM as non-invasive SFN marker; corneal nerves correlate with diabetic retinopathy

One-paragraph summary

Single-site observational cross-sectional study of 80 adults with type 2 diabetes mellitus (T2DM) at Alexandrovska Hospital, Sofia, comparing corneal confocal microscopy (CCM) parameters (corneal nerve fiber density CNFD, fiber length CNFL, branching density CNBD) against modified Neuropathy Disability Score, cardiovascular reflex tests, Sudoscan sudomotor function, vibration perception threshold, carotid intima-media thickness, body composition, and laboratory parameters. Headline findings: (a) corneal nerve parameters correlate negatively with diabetes duration, HbA1c, intima-media thickness, and vibration perception threshold; (b) patients with diabetic retinopathy showed significantly reduced CNFD and CNFL — establishing a direct corneal-nerve ↔ retinal-substrate linkage; (c) ROC analysis demonstrates significant discriminative ability of CNFD for detecting sudomotor dysfunction and of the HRV index for identifying peripheral neuropathy. For the FM project, this paper is a cross-condition methodology anchor for corneal CFM as a non-invasive SFN marker in the retinal-biomarker-diagnostic-priority panel — validating one of the panel's component modalities in a different disease context. The retinopathy ↔ corneal-nerve correlation is mechanistically important: it suggests the corneal-and-retinal substrate is a coupled compartment where neurodegenerative changes co-vary, supporting the project's framing of the retina as a multi-modal CNS-imaging substrate accessible non-invasively. The paper is bridging-tier for FM because it operates in T2D, not FM — but the methodology validation transfers.

Claims as triples

• corneal_nerve_fiber_density — predicts → small_fiber_neuropathy [evidence: ROC AUC for sudomotor dysfunction detection; confidence: established (in T2D)]
• corneal_nerve_fiber_density — correlates_with → retinal_substrate_changes [via diabetic retinopathy comparison; confidence: emerging]
• corneal_nerve_fiber_density — correlates_with → autonomic_dysregulation [CNFD ↔ HRV cross-correlation; confidence: emerging]
• hrv — predicts → small_fiber_neuropathy [HRV index discriminative AUC for peripheral neuropathy; confidence: emerging]
• corneal_nerve_fiber_density — absent_in → diabetic_retinopathy [reduced in retinopathy-positive subset; confidence: emerging — extends the corneal-retinal linkage claim]

Methods note

Single-center observational cross-sectional design, n=80 T2DM patients (age 18-75), recruited at Alexandrovska Hospital (Sofia). Multi-modal assessment: modified Neuropathy Disability Score; CCM-derived CNFD, CNFL, CNBD; cardiovascular reflex tests; Sudoscan for sudomotor function; vibration perception threshold; carotid intima-media thickness; body composition analysis; standard diabetes laboratory parameters. ROC analysis for discriminative-ability claims. No control group beyond within-cohort stratification (peripheral neuropathy 57.5%, autonomic neuropathy 66.7%).

Limitations

• T2D cohort, not FM. Direct FM relevance requires accepting the methodology transfer claim.
• No healthy-control group — within-cohort stratification limits causal claims.
• Single-site, single-country recruitment.
• Cross-sectional design — cannot establish temporal direction (does retinopathy cause corneal nerve loss, or do both reflect a common upstream substrate?).
• The retinopathy subset within the cohort is not characterized in detail (proliferative vs non-proliferative, severity stratification).
• Doesn't directly address whether corneal CFM can substitute for skin biopsy IENFD as a non-invasive SFN diagnostic — the comparison is implicit rather than direct.

Open questions raised

• Does the corneal-nerve ↔ retinal-substrate coupling observed in T2DM (diabetic retinopathy ↔ reduced CNFD/CNFL) extend to FM? (Direct retinal-biomarker priority extension — predicts that FM patients with reduced CNFD/CNFL should show concordant retinal-imaging changes.)
• Is corneal CFM CNFD reduction in FM concordant with skin biopsy IENFD reduction in the same patients, at the level Yordanova 2026 establishes in T2D? (Q-Re-6 in project open questions.)
• Do FM-autoimmune subset patients (anti-SGC IgG positive) show preferentially reduced corneal CFM parameters compared to non-autoimmune-subset patients?

Triangulation notes

• Methodology validation for retinal-biomarker-diagnostic-priority panel: corneal CFM is one of the four planned panel components (alongside OCT macular cube, OCT EDI, OCT angiography). Yordanova 2026 supplies cross-condition methodology validation showing CCM is robust enough for clinical-grade diagnostic application.
• Mechanistic linkage between corneal CFM and retinal substrate at the patient level (diabetic retinopathy ↔ reduced CNFD/CNFL). This linkage was implicit in the project's framing of the retina as a coupled CNS-imaging substrate; Yordanova 2026 supplies empirical demonstration in a non-FM cohort.
• Connects to Bandinelli 2025 (post-COVID SFN systematic review): both papers identify corneal CFM as an established SFN diagnostic modality across non-FM conditions. The cross-condition convergence supports the methodology's portability to FM.
• Connects to existing project SFN framework: small_fiber_neuropathy mechanism + ienfd skin-biopsy biomarker + corneal_nerve_fiber_density newly-added biomarker entity. Yordanova 2026 strengthens the corneal-CFM-as-IENFD-substitute claim.
• HRV-CCM concordance finding connects to the project's autonomic-dysregulation chain (Wirth & Scheibenbogen 2026, Azcue 2026) — supports the proposal that autonomic-subtype FM would show measurable corneal CFM changes.

Bridges

• B-Re-8 (Corneal CFM ↔ retinal substrate coupling, NEW 2026-05-23). One endpoint at corneal nerve density (CCM); other endpoint at retinal substrate (diabetic retinopathy in this paper; OCT RNFL/OCTA in FM context). Confidence: bridging — established in T2D, predicted to extend to FM. Closing: paired CCM + OCT measurement in FM patients with stratification by autoimmune / chain status.

Cure-path implications

The corneal-retinal coupling Yordanova 2026 demonstrates supports the project's framing of the ophthalmologic-imaging module in the biomarker-mapping cohort as a multi-modal multi-chain readout. Single 30-45-minute appointment yields corneal-nerve, retinal-nerve, retinal-microvascular, and choroidal-thickness measurements that are mechanistically coupled per the Yordanova evidence. The cost-effectiveness case for the panel strengthens.